Senior DevOps / Cloud Infrastructure Engineer — GCP + Cloudflare (Long-Term, Hands-On)

About the role We run a production SaaS / messaging platform (WhatsApp automation, CRM, and several customer-facing web apps) on Google Cloud Platform and Cloudflare. We're looking for one genuinely senior DevOps engineer to take full ownership of our engineering infrastructure and make it excellent — end to end. This is not a "learn on the job" or "ask an AI and paste the answer" role. We need someone who has personally operated this kind of stack in production and can reason about failure modes, blast radius, and cost without hand-holding. What you'll own - GCP: Compute Engine, Cloud SQL (PostgreSQL), Cloud Run, IAP-based SSH, VPC / firewall, Secret Manager, IAM / service accounts, startup scripts, billing & cost optimization - Cloudflare: DNS, Access (Zero Trust), Tunnel (cloudflared), AI Gateway, WAF / security rules - Reverse proxy & services: nginx (many vhosts / sites-enabled), systemd services, zero-downtime reloads with proper smoke testing across all domains - CI/CD & code health: GitHub (branch protection, CODEOWNERS, Actions CI), SonarCloud, Sentry, engineering metrics (Swarmia), Jira / Atlassian - Reliability & security: secret rotation, closing injection / XSS holes, observability, incident response, backups & rollback discipline - Cost optimization: right-sizing VMs and Cloud SQL, eliminating waste, forecasting spend — bring the bill down without breaking production You must have real, demonstrable experience with most of these GCP (Compute Engine, Cloud SQL, Cloud Run, IAM/IAP, Secret Manager) · Cloudflare (DNS, Access, Tunnel) · nginx · systemd / Linux (Debian/Ubuntu) · PostgreSQL operations · GitHub Actions CI/CD · Sentry · SonarCloud (or equivalent SAST) · Infrastructure-as-Code (Terraform a big plus) · Docker · shell scripting Please DO NOT apply if - You've only used these tools in tutorials or personal projects - Your plan is to figure it out by prompting an AI as you go - You've never operated a multi-VM, multi-domain production environment - You can't explain what happens when an nginx reload fails on a shared host To be considered, answer these in your proposal (generic proposals will be ignored) 1. Describe a production incident on GCP or Cloudflare that you personally diagnosed and fixed. What was the root cause and how did you find it? 2. How do you safely reload nginx across 10+ vhosts without risking the other domains? What's your rollback plan? 3. A Cloud SQL instance is 60% over-provisioned but hits CPU spikes 3× a day. How do you approach right-sizing? 4. How do you handle secrets that were accidentally committed to a git repo and are already live in production? 5. Roughly how many hours per week are you available, and what is your rate? Engagement Long-term and ongoing. We'll start with a small paid trial task so we can both evaluate fit. --- Posting tips (not part of the ad): - Put questions 1, 2, and 4 into Upwork's Screening Questions field too — those are the ones a copy-paste applicant can't fake. - Set experience level to Expert and type to Hourly (or Fixed for the trial). "Intermediate" attracts exactly the profile you're trying to avoid. - Never put real hostnames, IPs, project IDs, or API keys in the ad or the trial — use a dummy/isolated environment for the test task.